Teams of head and neck radiation oncologists, surgical oncologists and medical oncologists at Memorial Sloan Kettering Cancer Center (MSK) recently published two retrospective studies with practice-changing insights for improving the treatment and surveillance of patients with oropharyngeal cancer.
The first study, published online on November 26, 2025, in JAMA Otolaryngology-Head & Neck Surgery, found that the three-year rate of osteoradionecrosis after radiotherapy was low overall at 3%, but with a higher rate after proton therapy than intensity-modulated radiation therapy (IMRT), particularly in the definitive treatment setting. (1)
The second study proposes an optimized, risk-adapted post-treatment surveillance strategy for human papillomavirus (HPV)-positive oropharyngeal cancer. It demonstrated superior efficacy in boosting early relapse detection and greater cost-effectiveness than current approaches recommended by the National Comprehensive Cancer Network (NCCN) and the Radiation Therapy Oncology Group (RTOG). The paper was published online in Cell Reports Medicine on December 1, 2025. (2)
MSK treats hundreds of patients with oropharyngeal cancer annually, including both the New York Proton Cancer Center in New York City and the Procure Proton Center in New Jersey. All patients with head and neck cancer are captured in a prospective clinical database with detailed documentation of outcomes and toxicities. Both of these new studies evaluated data from patients with oropharyngeal cancer who were treated between January 2013 and December 2023.
“We prospectively collect clinical data for every patient so that we can conduct analyses to answer important questions with a view to improving patient outcomes,” said radiation oncologist and early drug development specialist Nancy Lee, MD, FASTRO, Director of Head and Neck Radiation Oncology and service chief of Proton Therapy at MSK, Vice Chairman of the Department of Radiation Oncology, and senior author of both studies. “Insights from these two studies are already informing clinical care and the design of future studies at MSK.”
First Study: High-Grade Osteoradionecrosis Rare After Radiotherapy with Some Differences Between Modalities
Study Design
Osteoradionecrosis (ORN) is a potentially debilitating late complication of radiotherapy for head and neck cancer. Proton therapy provides superior dose conformality compared to IMRT, but its impact on ORN risk in patients with oropharyngeal cancer has been unclear.
The MSK investigators reviewed medical records, clinical notes, and imaging results for 1,594 eligible patients with oropharyngeal cancer who received 50 Gray (Gy) or more of radiotherapy, and excluded patients with oral cavity cancer. They also cross-referenced lists with the MSK Dental Oncology Service, which maintains an active database of all patients with ORN. (1)
ORN was defined according to diagnostic criteria as the region of defect that received radiation and showed the presence of exposed bone in the irradiated field and/or an area of radiographic changes within the intact mucosa, (3) (4) (5) without evidence of residual or recurrent tumor at the ORN site. (3)
Patient Characteristics
The mean age of patients was 62 years. Of the total of 1,564 patients, 1,356 were male (87%), and 208 were female (13%). Most patients (89%) underwent IMRT and 11% received proton therapy. (1)
The IMRT cohort was slightly younger than the proton cohort, 61 versus 64 years, but there was no difference by sex. Among 1,555 patients for whom smoking status was available, 45% of the IMRT group and 58% of the proton group had never smoked. (1)
There were no differences between the IMRT and proton cohorts by disease subsite. However, there was a somewhat higher proportion of HPV-positive patients in the proton group (92%) than in the IMRT group (86%). The distribution of T-stage disease was similar between cohorts, but the IMRT group had a higher percentage of N2 or N3 disease. (1)
Most patients (84%) received 7,000 centi-Gy (cGY) in 35 fractions (200 cGy daily fractions for seven weeks). No patient had twice daily treatments. In the proton group, 83% of patients had concurrent chemotherapy, compared to 89% in the IMRT group. (1)
Study Findings
Among 1,564 patients in the overall cohort, 68 (3%) developed ORN of any grade during the study period, and among those who developed ORN, a majority had grade 2 events (74%), 12% had grade 1 events, and 15% experienced grade 3 events. Grade 3 ORN was rare at about 1% for both modalities. (1)
In the overall cohort, the three-year ORN rates varied by modality and patient characteristics as follows: (1)
- 6.4% after proton therapy versus 2.7% after IMRT, hazard ratio (HR) = 2.62.
- 3.2% for patients who received concurrent chemotherapy versus 1.9% for those who did not (HR = 3.16).
- 4.2% for those with a smoking history versus 1.7% for no smoking history (HR = 2.24).
- Proton therapy, concurrent chemotherapy, and smoking were independently associated with ORN, with HRs of 2.92, 3.29, and 2.33, respectively, on multivariate analysis.
- Radiation dose to the primary tumor as a continuous variable or modeled as a binary variable of more or less than 7,000 cGy was not associated with ORN, but wide confidence intervals prevented making a definitive conclusion.
Within the definitive treatment setting, the three-year ORN rate was higher with proton therapy at 7.47% than with IMRT (2.38%; HR = 3.62). (1)
Study Implications
“It’s encouraging to find that advanced ORN remained uncommon with both modalities in a relatively homogeneous cohort of patients with oropharyngeal cancer across a decade,” said Dr. Lee. “However, this complication continues to occur, and methods to mitigate ORN are needed.”
“Given these results, we are now exploring a modified two-stage approach to radiotherapy for OPSCC at MSK, using IMRT followed by proton therapy to mitigate the risk of ORN, especially among patients treated in the definitive setting,” she said. “Future research will be required to quantify the benefit, especially among patients treated in the definitive setting.”
The study was supported by the National Institutes of Health/National Cancer Institute grant P30 CA008748. Access disclosures for Dr. Lee.
Second Study: A Personalized Risk-Adapted Surveillance Strategy Boosts Early Relapse Detection and is Superior to Guidelines for Efficacy and Cost-Effectiveness
Early detection of recurrence through post-treatment surveillance is critical for enhancing survival outcomes in patients with oropharyngeal cancer, as many early recurrences can be addressed with potentially curative treatment.
However, there is no current consensus on the optimal surveillance schedule and regimen. Existing recommendations from the NCCN and RTOG differ greatly on follow-ups for history and physical examinations as well as imaging protocols. (6) (7) Further, current strategies are applied to all patients with HPV-positive disease, regardless of heterogeneous patterns of recurrence across different disease stages.
Study Design
Dr. Lee and colleagues evaluated a large cohort of 1,146 HPV-positive oropharyngeal cancer patients who had undergone IMRT. Using a random survival forest model, they quantified the monthly risk of disease failure with high resolution. (2)
Next, they developed a stage-specific, risk-based surveillance strategy designed to optimize recurrence detection at each follow-up visit. Finally, they compared their proposed risk-based surveillance strategy with existing clinical guidelines, including NCCN27, NCCN15, NCCN10, and RTOG. (2)
Patient Characteristics
Of the total cohort of 1,146 patients, 55.5% (636) had stage 1 disease, 26.1% (211) had stage 2 disease, and 18.4% (299) had stage 3 disease. (2)
The median duration ot follow-up was 63.4 months. Imaging at routine examinations included 575 magnetic resonance imaging scans of the neck, 1,170 neck computed tomography (CT) scans, 1,286 chest CT scans, and 2,626 positron emission tomography/CT (PET/CT) scans. (2)
A total of 34 patients (3%) experienced local recurrence, 41 (3.6%) developed regional recurrence, and 109 (9.5%) experienced distant metastasis. The median times to these events were 17.3, 13.7, and 11.6 months, respectively. (2)
Proposed Surveillance Strategies
For each disease stage, the investigators determined the optimal schedule for follow-up visits based on the distribution of risk-adjusted disease failure probabilities. The majority of follow-up visits were concentrated in the first two years after treatment. The proposed new risk-based surveillance schedule for years 1 through 5 is as follows: 10 visits for patients with stage 1, 11 visits for those with stage 2, and 12 visits for patients with stage 3 disease. Refer to Figure 3 in the paper to see how these visits are distributed. (2)
Study Results
The proposed surveillance strategy was more effective and more cost-efficient than guidelines-recommended strategies. (2)
The new risk-based surveillance strategies by disease stage consistently detected recurrence earlier than current guidelines. This finding was particularly true for patients with stage 1 disease: the proposed plan with a lower number of follow-up visits demonstrated superior disease detection compared to NCCN27, which involves 27 visits across five years. The benefit of risk-based surveillance was less pronounced in patients with stage 3 disease: the risk-based schedule of 12 visits resulted in similar detection performance as NCCN15 and RTOG over five years. (2)
In terms of cost-effectiveness, the proposed risk-based surveillance strategy was superior to the control strategies across all disease stages. For example, in patients with stage 1 oropharyngeal cancer, the incremental cost-effectiveness ratios (ICERs) were as follows: (2)
- NCCN15: $48,431 per quality-adjusted life years (QALY)
- NCCN27: $103,501 per QALY
- RTOG: $44,049 per QALY
- Proposed risk-based strategy: -$39,900 per QALY
The risk-based strategy was also the most cost-effective option compared to existing guidelines for patients with stage 2 and stage 3 disease, with ICERs of -$23,055 per QALY and -$317 per QALY, respectively. (2)
Study Implications
“Our proposed stage-specific, risk-adapted surveillance schedule for optimizing the detection of clinically significant events per visit has transformed care at MSK,” said Dr. Lee. “These personalized strategies will result in improved patient outcomes at a more effective cost.”
“On a final note, keep in mind that our study included a large, real-world cohort of patients with oropharyngeal cancer treated at a single, high-volume tertiary care center where five-year disease recurrence is exceedingly rare,” Dr. Lee said. “Further external validation with independent datasets is needed before generalizing our results to other centers.”
The study was funded in part through the National Institutes of Health/National Cancer Institute support grant P30 CA008748. Access disclosures for Dr. Lee.